1. review existing evidence that the introduction of pest resistance into a crop species has affected the establishment, persistence, and spread of the crop or sexually compatible species; and
2. identify, and recommend research strategies to address, gaps in information concerning the effects of pest resistance genes on the establishment, persistence, and spread of a cultivated crop or sexually compatible weedy species.

The focus throughout was to support and promote sound decision making by those who set research priorities, plan breeding programs, make regulatory decisions, or address public concerns regarding the use of genetically engineered crops.

In their invited talks, plenary speakers gave overviews and insights into crop breeding, weeds, pest resistance, ecology, and regulatory concepts. This material provided the background and context for the discussions that followed. Participants were invited on the basis of their expertise and with an eye towards achieving a balance of disciplines and institutional affiliations. They were organized into small multidisciplinary working groups, each centered around one of the seven target crop groups.

Group leaders, in consultation with experts, collected background information on their crop group and sent it in advance to participants. Topics included:

At the workshop, the groups were asked to assess what is known using the following guidance questions as a framework for their discussions:

• What is the evidence that introduction of a pest resistance trait could increase the ability of the crop to become established, persist, or spread?
• What is the evidence that pests have a significant effect on populations of plant species that are sexually compatible with the crop? Are any such pests common to both the crop and the sexually compatible species?
• If the crop is made resistant to such pests, what are the potential consequences of the pest resistance trait moving from the crop to the sexually compatible relatives? How likely is introgression of the resistance trait?

The groups were then asked to identify what is needed for sound decision making by considering:

• What specific information is not currently available but would be important in providing a stronger scientific basis for evaluating the effect of pest resistance genes on the establishment, persistence, and spread of the crop or its sexually compatible relatives?
• What are the available sources and/or experimental approaches that would provide such needed information?
• What characteristics of the crop affect our ability to extrapolate from small-scale field tests to large-scale use in terms of evaluating its establishment, persistence, and spread?

The following summaries highlight the main conclusions and recommendations emerging from the group discussions. The full reports, which integrate the working groups’ collective knowledge and insight, should be a valuable resource for persons involved in making decisions on the appropriate development and use of pest resistant varieties of these crops.

Strawberry (Fragaria spp.), blackberry and raspberry (Rubus spp.), and blueberry (Vaccinium spp.) are small berry crops with potential to hybridize with feral populations of weedy relatives. Strawberries are known to escape from cultivation and to cross with wild relatives, but generally lack aggressive weedy characteristics. Although cultivated strawberries are subject to attack by a wide variety of diseases, little evidence of disease has been noted on leaves or fruit of wild populations, supporting the assumption that environment is a greater limiting factor than pest pressure on strawberry establishment. Nevertheless, important information needed to verify this hypothesis is missing. Data is needed on the ecology of wild strawberry populations, the level of hybridization between crop and wild relatives, and the impact of pests on wild populations.

Most of the pest resistance traits incorporated into blackberry and raspberry have been derived from weedy relatives. As a result, there is currently little concern about escape of pest resistance genes to wild relatives. Furthermore, there is no evidence that resistance traits bred into raspberries over the past 60 years have increased the weediness of the crop. In contrast, the working group felt that engineered herbicide tolerance in Rubus would be unwise because it was likely to confer a selective advantage on weedy relatives in agricultural settings and eliminate important weed control options. Important information for the risk assessment of pest resistance genes should include surveys of pest incidence on weedy species.

Introgression between cultivated and wild blueberries has been documented, but the group did not think the impact of pest resistance genes on either the crop or relatives was a matter of serious concern. Neither cultivated nor wild blueberries have characteristics associated with aggressive weeds, and the transfer of a single pest resistance trait was not seen as likely to alter this.

Discussion focused on the common Brassica species that are currently most subject to modification by genetic engineering, the oilseed crops B. napus and to a lesser extent, B. rapa and B. juncea. All are capable of cross hybridizing among themselves and with other related species. Many commercial cultivars already contain resistance to common fungal pathogens; because the resistance was derived from wild relatives, it was considered unlikely that movement of such genes back to weedy relatives would have a significant impact on fitness of the weeds. However, the situation may be different for other genes encoding traits such as insect resistance and herbicide tolerance (or most problematic, a combination of both), which could confer a substantial fitness advantage on a recipient plant.

Available information indicates that cultivated transgenic Brassica will hybridize with a number of weedy species and that introgression of transgenes is probable. Ecological studies show that in many environments insects are the principal factor limiting plant population growth, suggesting that acquired pest resistance genes could increase the fitness and hence the population range of weedy Brassica species. However, too little information is available to definitively state that this risk would outweigh the benefits of having crops with enhanced pest resistance.

The working group identified seven areas of research that would contribute to our knowledge of pest resistance gene impact on Brassica species:

1. The creation of a database of sexually compatible species and varieties.
2. The development of a geographic information system of pest influence. This would combine species ranges with environmental information required to predict the impact of pests on a given host.
3. Long-term studies on weed populations to examine changes in pest resistance gene frequencies and the effect of such changes on pest populations.
4. Pest exclusion studies to measure the influence of pest pressure on plant reproductive rates.
5. Hybridization and introgression experiments using resistance-conferring transgenes to measure the performance and persistence of transgenes in the environment.
6. Observational studies of basic reproductive biology of lesser-studied related species.
7. Modeling projects to synthesize available knowledge and direct future research.

Sorghum (Sorghum bicolor), rice (Oryza sativa), and wheat (Triticum aestivum) have close weedy relatives capable of hybridizing with the respective cultivated crops, although the ease of introgression depends on the specific crop-weed complex. Pest resistant varieties of these crops are currently being bred using traditional and (except for sorghum) genetic engineering techniques. The group could find no evidence that pest resistance traits introduced into these crops or their weedy relatives would affect their ability to establish, persist, or spread. Because these crops (and associated weeds) are already subject to integrated weed management programs, there was little concern about exacerbation of a weed problem within the managed agroecosystem. However, insufficient data exists to make the same conclusion about less managed ecosystems.

The working group concluded that pest resistance genes derived from the same gene pool (i.e., characterized and predictable genes from conventional breeding or genomics programs) are of low risk. More information is needed to assess the risk of introducing genes derived from diverse sources. Recommended research topics on the environmental effects of novel genes introduced into crop species include:

1. Inventories of pest infestations of the related weedy species.
2. Presence of pest resistance traits in the weed population.
3. Impact of pests on weed population dynamics in the absence of resistance.
4. Where impact is significant, quantitation of pest infestation.
5. Studies of crop-weed hybrids if fitness or population dynamics is affected.

When an engineered pest resistant crop deemed to present low risk based on small scale studies, is released commercially, the first five years following release provide a unique opportunity for risk assessment on a larger scale. It was recommended that funding and research efforts be targeted to this time period.

The diversity in origin and genetic composition of wild and weedy cucurbit crops makes generalizations about crop-weed complexes difficult. Cucurbits grown in the US have both Old World and New World origins and are generally interfertile with wild native or introduced cucurbit species in the US. Weedy relatives include dudaim (Cucumis melo subsp. melo) and varieties of Cucurbita pepo. Evidence indicates that wild C. pepo has experienced hybridization and introgression with cultivated relatives, and perhaps that some weedy species have evolved from escaped ornamental gourd varieties.

It is possible that introduced pest resistance traits in Cucurbita spp. could enhance weediness, but this would depend on whether the trait conferred a selective advantage on the recipient plant. An introduced gene for virus resistance has been demonstrated to flow from squash to wild relatives and confer virus resistance to the wild plants. Although viruses have been reported on wild C. pepo, the impact of viruses on such wild populations has not been investigated.

The group concluded that our ability to evaluate the risk of pest resistance genes enhancing a weed problem would benefit from a greater understanding of the biology of wild cucurbit species. This would include weediness characteristics (i.e., degree of aggressiveness), genetic similarity of crop and weed, geographical distribution, ecological requirements, sympatry (degree of genetic interaction among crop and related weed), reproductive biology, pests of wild species, and pest resistance in wild species (i.e., type, frequency, stability). Because of the release of transgenic virus resistant summer squash, efforts should focus on wild C. pepo as well as other weedy Citrullus and Cucumis species.

The turfgrasses creeping bentgrass (Agrostis palustris) and Kentucky bluegrass (Poa pratensis) were considered by the group because of recent efforts to genetically engineer these crops. Turfgrasses are highly domesticated species that are subject to intense management. Although they are capable of hybridizing with wild relatives, they are relatively slow growing, small, and quickly out-competed by most plants. These traits, combined with the fact that mowing normally prevents these plants from setting seed, suggests that crop to weed gene flow and introgression would be rare.

Various disease and insect resistance traits have been bred into turfgrasses, and these have contributed to an expanded geographic range of cultivation. However, the working group was unaware of any evidence that introduction of a pest resistance trait had resulted in turfgrasses or their sexually compatible relatives overcoming any control exerted by those pests.

Despite the low weediness potential of turfgrasses, the group identified several gaps in our knowledge that could be filled by research on:

1. The life history and invasiveness of the various turfgrass species.
2. The geographic range of related species, as well as their cross-compatibility with crop species.
3. The diversity of pests and pathogens that attack the sexually compatible relatives.
4. The factors (including pests and pathogens) that limit populations of sexually compatible relatives.
5. The rate of increase of populations of sexually compatible relatives, and the factors that control them.
6. A greater understanding of the characteristics of weedy grasses in general.

It was proposed that this information could be obtained through several avenues, including existing literature, from which information could be compiled into a useful database; introduction experiments, in which transgenic plants could be monitored after controlled introduction into populations of sexually compatible wild relatives; simulation experiments, which simulate greater reproductive fitness by artificially increasing seed output in target plant populations; and experimental crosses, which directly characterize the weediness potential of hybrid progeny of transgenic crops and weedy relatives.

Poplar (Populus spp. including cottonwoods, aspens, and related hybrids) differ considerably in their biology from other crop groups considered in the workshop. They become reproductively active between the ages of five to fifteen years, have long life spans, and exhibit a high capacity for vegetative regeneration. Dispersal by sexual reproduction can be extensive whether pollen or seed is considered, however, seeds rapidly lose viability and thus do not persist in the seed bank. Many species have stringent habitat requirements (e.g., for moisture), most environments harbor large wild populations compared to poplar plantations, and seedlings are not competitive in closed stands of trees or herbs. This creates a condition of "genetic inertia" in which significant changes in population genetics due to transgenes may take dozens to hundreds of years to occur. Nevertheless, poplars are very closely related to their wild relatives, and gene flow and introgression have been documented.

Development of poplar with fungal resistance is proceeding primarily by conventional breeding, but genetic engineering of insect resistance (and herbicide tolerance) is well under way. The ability of poplars to disperse pollen and seeds over great distances suggests that transgenes from plantation trees will escape, but the genetic inertia of these trees makes it difficult to predict when, or even if, transgenes might significantly impact wild populations. Despite the large areas and long time required to study this issue, the working group concluded that the risks of releasing transgenic poplar do not outweigh the benefits, but that releases should be coordinated with monitoring programs to follow the impact of these genes on the environment.

Seven areas were identified where research would be useful to inform both scientific and regulatory decisions on pest resistance genes in poplar. Ranked from highest to lowest priority, these are:

1. Isolation of additional kinds of insect and disease resistance genes.
2. Development of reliable containment methods to prevent seed and pollen movement of transgenes (e.g., engineered tree sterility).
3. Information to support management of pest resistance (e.g., insect dispersal, refugia design).
4. Poplar reproductive biology, seed, and pollen dispersal, and fertility of crop-wild hybrids.
5. Ecology of natural pest resistance mechanisms in relation to species interactions and ecosystem function in the wild.
6. Evaluation of the economic and legal impacts of transgene spread.
7. Analysis of contributions that transgenic poplars could make to economic and environmental sustainability.

Cultivated sunflower (Helianthus annuus) overlaps in range with its weedy, wild progenitor (also H. annuus) and the two are fully capable of hybridizing. Pollen from cultivated sunflower may be spread to adjacent wild populations through the movement of insects, and thus crop genes may introgress and persist in populations of wild sunflower.

Disease resistance in cultivated sunflower has been obtained through both conventional and transgenic approaches. Insect predation on cultivated sunflower is considered to reduce yield both by directly consuming seed heads or by spreading disease agents, thus engineering resistance to insects (i.e., Bt toxin) is a high priority for transgenic commercial hybrids. No studies have examined whether gene flow from cultivated to wild sunflower has had an impact on the wild population. Since the pest resistance traits bred into crop cultivars to date have largely been derived from wild germplasm, it is not likely that these traits would add anything new to the wild populations.

The working group developed a series of questions that should be addressed for each new type of transgene that confers resistance to insects or disease. They provide a framework for identifying important research areas and aid in making decisions about the release of transgenic sunflower.

1. Is the transgene inherited as a stable, Mendelian trait when it is crossed into wild plants? Experiments to address this question should examine genetic behavior of the new trait and its expression under various environmental conditions.
2. Do insects or diseases targeted by the transgene occur in populations of wild sunflower, and if so, how common are they? A recommended approach is to conduct detailed surveys to examine the influence of targeted pests on wild sunflower populations.
3. When the transgene has introgressed into wild plants, will these plants exhibit greater survival or fecundity than their nontransgenic counterparts? Suitable experiments would examine the impact of pest resistance genes either through simulation or controlled introgression of the gene into wild relatives.
4. If the transgene leads to greater survival or fecundity, will this cause wild populations to become more troublesome weeds? A combination of field experiments and modeling to predict potential impacts could provide important insights.

Despite the diversity of participants’ backgrounds and the range of crops discussed, there were several points of general consensus. The common baseline was a recognition that conventional agricultural activity entails certain environmental and ecological risks. Given that, the group concluded that the genetically engineered pest resistance traits currently being field tested or commercially released present no fundamental differences from similar traits bred into crops using traditional techniques. It should be noted that some participants disagreed, however, and contended that transgenes will have more profound effects on crop phenotype than traditional genes, and thus potentially greater impact on weed species.

The second point of consensus was the view that cases in which crops are engineered with multiple pest resistance or other fitness traits present more complex ecological questions. Such "gene stacking" to confer resistance against a broad spectrum of pests may give recipient plants a greater selective advantage and lead to ecological consequences that are less predictable than the single-gene pest resistance traits which constitute much of our experience to date. Participants agreed that the general consensus on the nature of risks posed by current transgenic crops could not be extended to the next generation of crops engineered with multiple pest resistance traits.

Organization of the working groups around crop types proved to be a very effective approach for synthesizing what is known and what needs to be known about ecological effects of introduced pest resistance genes. Although all groups were given an identical set of guidance questions, each group struggled with a unique crop-weed situation and set of issues. The biology and ecology of the crop and its weedy relatives dictated which issues were most relevant, and was perhaps the most important factor in determining the groups' level of concern over the risk of transgene escape. The weediness of the crop and its wild relative, the probability of crop/weed hybridization and transgene introgression, the life spans of the crop and weed, and the persistence of each in the seed bank all influenced the groups’ thinking. Other important parameters that varied by crop group were the susceptibility of crops and weeds to pathogens and the role such pathogens play in limiting populations.

Discussions during the workshop were based on information compiled by group leaders prior to the meeting and the knowledge and experience participants brought to the table. Within every working group, members agreed that much information is lacking about the ecology of crop-weed complexes, in particular the level to which pests limit weed populations. This shortcoming hampered their ability to accurately predict the consequences of novel pest resistance traits. More importantly, it resulted in specific recommendations for more research on basic plant biology and ecology, as well as applied risk assessment. Although most commercially important crops have related weedy species somewhere in the world, not all of these crops are expected to be engineered for pest resistance in the near future. It is therefore a feasible task to generate essential biological and ecological information on the more widespread outcrossing crop species, which would increase our ability to make educated determinations of risk posed by release of genetically engineered varieties.

Table of Contents